I'll try to take a look this week-end (starting this evening).
You can also ty to see what is the asm generated from your code to have an overview of what seems to be too complicated compared to your C code.
Without rewriting parts in asm, you can try to modify your c code and see if the generated asm is better (more simple) or worse.
For example, doing multiply in loops could be replaced by cumulative additions...
You can also check with Tixul if he can add temetry in his emu...
You can also ty to see what is the asm generated from your code to have an overview of what seems to be too complicated compared to your C code.
Without rewriting parts in asm, you can try to modify your c code and see if the generated asm is better (more simple) or worse.
For example, doing multiply in loops could be replaced by cumulative additions...
You can also check with Tixul if he can add temetry in his emu...